← work

A model-agnostic knowledge base that onboards engineers onto complex service repos

Client
A web-data infrastructure company
Industry
Data infrastructure
Year
2025
Role
Forward-deployed engineering
  • Node
  • Python
  • Redis
  • MongoDB
  • Kubernetes
  • RabbitMQ
  • PostgreSQL
  • Knowledge systems
  • AI-native tooling
  • Developer onboarding

A model-agnostic company knowledge base — works with Claude, Codex, or Copilot — that captures senior engineers' system reasoning and onboards new hires onto complex service repos. Adopted and actively contributed to by the team.

Context

Embedded with a distributed engineering team where critical system knowledge lived siloed in individual engineers’ heads, and new hires struggled to onboard.

Approach

Interviewed the engineers to extract the reasoning behind their systems and built a model-agnostic knowledge base — one that works with Claude, Codex, or Copilot.

Outcome

The knowledge base now onboards new hires onto complex service repos with minimal friction, and is adopted and actively contributed to by the team. Separately, scraping success rates improved across major target domains.